Glossary term

Resolution Bandwidth

Engineering definition of resolution bandwidth covering RBW, spectrum analyzer filters, FFT bin settings, displayed noise, spectral separation and validation.

Branch: Telecommunications Engineering
Glossary type: metric
Content: Glossary term
Updated: Jun 26, 2026
Revision: v1.0.0 · reviewed

Definition

metric

Resolution bandwidth is the spectrum-measurement filter or bin bandwidth that controls how closely spaced frequency components can be separated and how much noise is displayed per trace point.

Resolution bandwidth, often abbreviated RBW, is used in spectrum analyzers, monitoring receivers and FFT-based measurements. Narrower RBW can reveal closely spaced carriers and reduce displayed broadband noise, but it increases sweep or record-time requirements and may miss short events. Wider RBW is faster and useful for overview scans, but it can hide narrow signals and raise the displayed noise floor.

Resolution bandwidth, usually abbreviated RBW, is the bandwidth of the measurement filter or spectral bin used to form a spectrum trace. It affects two things engineers often confuse: frequency separation and displayed broadband noise. A narrower RBW can separate closer tones and lower the displayed noise per point. A wider RBW can scan faster and catch wider emissions, but it may merge narrow signals and raise the apparent noise floor.

RBW is not just a display preference. It is part of the measurement evidence. A spectrum occupancy plot, spectral mask result, adjacent-channel leakage check or interference survey without RBW is incomplete because changing RBW can change whether a signal is visible, whether two signals are separable and where the trace sits relative to a threshold.

Measurement Filter

For a swept spectrum analyzer, RBW is the bandwidth of the intermediate-frequency or digital resolution filter. For an FFT-based analyzer, it is related to bin spacing, window shape and equivalent noise bandwidth. A simplified bin-spacing relation is:

\displaystyle \Delta f=\frac{f_s}{N}

where f_s is sample rate and N is record length. The effective measurement bandwidth also depends on the window:

B_{meas}\approx B_{ENBW,bins}\Delta f

so bin spacing and noise-equivalent bandwidth are related but not identical.

Noise Scaling

If broadband noise density at the measurement reference plane is N_0 in dBm/Hz, displayed integrated noise over RBW can be screened as:

N_{RBW}=N_0+10\log_{10}(B_{RBW})+C

where C is any analyzer or window correction needed to align the displayed bandwidth with the noise-equivalent bandwidth. If two RBW settings use the same correction convention, the displayed noise change is:

\displaystyle \Delta N=10\log_{10}\left(\frac{B_{RBW,2}}{B_{RBW,1}}\right)

This is why increasing RBW by a factor of ten raises the displayed broadband noise by about 10 dB.

Worked Example

A site survey records an average broadband noise density:

N_0=-154\ \text{dBm/Hz}

With:

B_{RBW,1}=10\,000\ \text{Hz}

the displayed noise is:

N_1=-154+10\log_{10}(10000)=-114\ \text{dBm}

With:

B_{RBW,2}=100\,000\ \text{Hz}

the displayed noise is:

N_2=-154+10\log_{10}(100000)=-104\ \text{dBm}

The environment did not become noisier. The measurement admitted ten times more broadband noise into each displayed point.

Now suppose two narrow carriers are separated by:

\Delta f_{tones}=18\ \text{kHz}

and the review rule requires at least 1.5 RBW of separation:

M_{res}=\Delta f_{tones}-1.5B_{RBW}

With 10 kHz RBW:

M_{res}=18-1.5(10)=3\ \text{kHz}

The tones are barely resolvable. With 30 kHz RBW:

M_{res}=18-1.5(30)=-27\ \text{kHz}

they will likely merge into one displayed feature.

Boundary With ENBW and Bandwidth

Bandwidth is the broad system concept. Equivalent noise bandwidth is the rectangular bandwidth that passes the same white-noise power as a real filter or window. Resolution bandwidth is the chosen spectrum-measurement bandwidth. In a calibrated instrument these concepts may be internally corrected, but a report still needs the RBW setting because it controls repeatability and interpretation.

RBW also differs from video bandwidth. Video bandwidth smooths or filters the detected trace after the spectrum measurement. It can reduce visual variation but does not create the same frequency resolution as a narrower RBW.

Validation Evidence

A defensible spectrum measurement states center frequency, span, RBW, video bandwidth if used, detector mode, averaging, sweep time or record length, window, reference level, attenuation, preamplifier state, antenna or conducted boundary, cable loss, calibration state, timestamp and uncertainty allowance.

For intermittent signals, RBW must be coordinated with sweep time and detector mode. A narrow RBW with a slow sweep may miss short bursts. A wide RBW with peak hold may reveal the burst but obscure close frequency structure.

This tradeoff should be chosen from the decision. A regulatory spectral-mask test usually needs the specified RBW and detector. A field interference hunt may need several passes: a wide, fast scan to find intermittent activity, followed by a narrower RBW capture to separate carriers and estimate occupied frequency. Treating one trace as universal evidence is rarely defensible.

Common Mistakes

Common mistakes include comparing traces with different RBW settings, treating lower displayed noise as a real environmental improvement, using wide RBW for narrow interferer searches, using narrow RBW for fast burst detection without checking sweep time, omitting detector mode, and confusing RBW with occupied bandwidth or channel bandwidth.

The practical rule is to record RBW wherever a spectrum trace supports an engineering decision.

REF

Disciplines